ETF Expert Corner

EquBot Co-Founder Spotlights World’s First Artificial Intelligence ETF

February 27th, 2018 by ETF Store Staff

Art Amador, Co-Founder of EquBot, spotlights the AI Powered Equity ETF (AIEQ).



Transcript

You can listen to our interview with Art Amador by using the above media player or enjoy a full transcription of the interview below.

Nate Geraci: The ETF we're spotlighting this week is the AI Powered Equity ETF, ticker symbol AIEQ. Joining us via phone from San Francisco to discuss this ETF is Art Amador, Co-Founder of EquBot. EquBot developed the quantitative model that powers this ETF. That model runs on IBM Watson artificial intelligence. EquBot is also the ETF sponsor. Art, our pleasure to have you on the program today.

Art Amador: Hi Nate. Thanks for having me.

Nate Geraci: Art, first tell us a little bit about EquBot and then I'd love to hear the backstory on where the idea for this ETF came from.

Art Amador: Yeah, sure, absolutely. EquBot is a financial technology company, as you said, based here in San Francisco. We are applying artificial intelligence to analyze investment opportunities. Currently, we are using AI to analyze over 15,000 global companies. The idea of EquBot actually started during a classroom discussion at the Haas School of Business at UC Berkeley. Also, as you mentioned, EquBot is an IBM Global Entrepreneur with Watson company. Anything else I can answer about EquBot before I jump into the story?

Nate Geraci: No, I'd love to hear the backstory.

Art Amador: Okay. One day at UC Berkeley I was discussing with Chida Khatua, who is the CEO and Co-Founder of EquBot, my clients’ high expectations. So before this, I worked for Fidelity Investments, working with high net worth investors. I was responsible for over about a billion dollars of high net worth assets. I was explaining to him that my clients had very high expectations. They expected me to know everything about markets and companies. At the same time, Chida was taking an investment class where prominent investors would come in and discuss their deep expertise in one specific market niche. It was kind of then when Chida said we could utilize artificial intelligence to analyze markets and companies and go beyond what an army of research analysts could cover. Furthermore, he believed as the amount of data increases, so does the level of decision making when it comes to investing. The company, we started experimenting with IBM Watson back in the summer of 2016. The company was formed in January of 2017 and here we are today with the AI Powered Equity ETF managing about 150 million dollars in assets. It's been a very interesting ride. We are very humbled by the interest that we've seen from investors.

Nate Geraci: Art, before we get into the details of the ETF, for listeners who aren't familiar with IBM's Watson artificial intelligence - just at a high level - what is this and how does it work?

Art Amador: IBM's Watson AI is a computing platform, or artificial intelligence system, that's capable of answering natural language questions by connecting large amounts of data - both structured and unstructured data. And it learns from each analysis that it conducts so that it can produce a more accurate answer with each subsequent question. The way to think about this, and it's the same way we think about investment strategies, the value of the system actually grows over time as it continues to learn and make more and more intelligent decisions. Let me clarify a couple of things. I used the words "structured" and "unstructured" data. IBM analyzes structured data. These would be things like spreadsheets. Unstructured data would be, for example, articles in the New York Times or Wall Street Journal. So IBM has a lot on the table to actually read through all those different documents and draw insights out of them.

Nate Geraci: So Art, who actually decides what data is fed into the computer? Is that data prioritized in any way?

Art Amador: The answer is a little bit more complex, but let me try to make it a little bit more simple. The data that we're collecting is both data that EquBot is responsible for going out and collecting, and then IBM's Watson also already has a bunch of data already collected. We're actually using both. A little bit more on the data side. We use both paid and unpaid data. What that looks like, you can think of paid data as something like a subscription to Bloomberg data, right. Unpaid sources would be, for example, the New York Times or Wall Street Journal or something to that extent where EquBot has its own web crawler that goes out and finds that information. As I mentioned, there's a data body that IBM has as well. When we’re asking the system questions, it's pulling in information from all those different pieces.

Nate Geraci: Our guest is Art Amador, Co-Founder of EquBot. Art, let's talk about the ETF itself. Again, the ETF is the AI Powered Equity ETF, ticker symbol AIEQ. Walk us through how this ETF is constructed.

Art Amador: Yes, the high-level concept here is that the AI technology mimics the investment process of an army of equity research analysts that are working around the clock. We say an army of equity research analysts because it's really the way in which an investor does deep due diligence on a company. We're utilizing publicly available financial data, such as 10Ks and 10Qs and current market forecast data to build predictive financial models on over 6,000 US publicly traded companies. We're also analyzing over a million news articles and social media postings every day. We use that information to analyze sentiment on the companies, sentiment on the market, to analyze the effectiveness of the management team, and to identify events or catalysts that are going to push the undervalued companies closer to their true intrinsic value. The way to think about how the portfolio is constructed, it does analyze 6,000 US companies, so it is completely domestic-oriented. It is a long only strategy. And collectively, there is about 70 positions in the portfolio. Now, the prospectus highlights that, in aggregate, the goal is to match the overall volatility of the broader US market, while utilizing those 70 positions. Think of it as an optimization problem. What you are ultimately doing here is you're looking at the 6,000 companies, you're seeing which ones are undervalued, what's happening out there within the news, and which ones have the highest potential for opportunity. But the way in which they're weighted takes into consideration the volatility of them individually and their correlation between the 70 companies. In aggregate, the risk, measured by standard deviation, matches the broader market but ultimately those 70 aggregates have the highest opportunity for capital appreciation.

Nate Geraci: Art, this is an actively managed ETF, but I'm curious, what exactly does that mean? How much human involvement is there given that the ETF relies on Watson artificial intelligence?

Art Amador: The EquBot investment process is as autonomous as possible. So from gathering data to generating ideas, to building a portfolio, and finally to monitoring that portfolio, that is all done through the AI system. The process is actually autonomous by design because one of the benefits that we believe is that it actually removes significant human bias and errors that some of the other traditional strategies fall victim to. However, I do want to point out though that the actual implementation of the trades are being done by people.

Nate Geraci: And how often does the Watson computer recommend a trade in the underlying holdings of this ETF? What does the turnover look like?

Art Amador: It is completely data dependent. So the data will kind of determine how much turnover actually happens. But as you mentioned, it is an active strategy so the expectation should be that it will trade more often than not. So right now, there has been – we launched October 17th - there has been a handful of days where it hasn't traded, but most days it does trade. You're looking at, probably on a daily basis, about a 1% turnover. But as I said, it is completely data dependent so the expectation should be that.

Nate Geraci: Our guest today is Art Amador, Co-Founder of EquBot. We’re spotlighting the AI Powered Equity ETF, ticker symbol AIEQ. Art, we always like to take a full 360-degree view of any ETF we spotlight. One of the criticisms I've seen of using artificial intelligence in investing is that markets are, in large part, based on investor psychology. And of course, investors aren't always rational. So if you have a highly rational AI program and highly irrational investors, that could potentially be a problem for AI driven strategies. Any thoughts on that?

Art Amador: Yes, so behavioral finance is an area of interest of ours. As I've mentioned, we already talked about human bias, so how to determine if market behavior is rational or is irrational is something that we are currently working on. Currently, our models that do track some of the market behavior - and it’s been trained both in back testing and during live testing - during this process, our model has established some broad correlations on the market volatility. However, as I mentioned, this is an area of interest of ours and we do think that -we haven't done this yet - we do believe that measuring how rational/irrational a market is at any given time and which way it is moving is possible. So hopefully in a future conversation we can talk more about that.

Nate Geraci: So Art, for investors considering this ETF, where does it fit in a portfolio? Do you view this as a core holding?

Art Amador: Actually, we do view this as a core holding. It could be a replacement both for a S&P 500 fund or even a Russell 2000. It is dynamic in nature so it does change. However, if this is your first time being introduced to an AI product, then I would encourage you to at least try it out - consider it a new part of your strategic alpha positioning within your portfolio. Ideally, and over time, we will prove that this absolutely should be a core holding in your portfolio.

Nate Geraci: I think you began to allude to this earlier, is this ETF benchmarked against any particular index? Just given that it can hold any US stock, I'm just wondering how you'll evaluate performance. What's the best benchmark here?

Art Amador: So, officially the fund doesn't have a benchmark and identifying the appropriate benchmark is a little difficult because of the dynamic nature of the strategy. The fund did initially start out with the average company being a small or microcap company and now the average company is more of a mid-cap company. I think if you're going to benchmark this against something, it's probably best to use two benchmarks. One would be the S&P 500 and the other would be the Russell 2000.

Nate Geraci: Art, we have just a few minutes left before we let you go, clearly there has been a lot of interest in this ETF. As you mentioned, it launched back in October and already it has nearly 150 million dollars invested in it. Given that success so far, how do you see this area of ETFs evolving moving forward? Do you expect to see other AI-driven ETFs? Is EquBot considering other AI ETF strategies? What will the future look like here?

Art Amador: Yeah, absolutely. Let me first talk about EquBot's future plans here. As I mentioned, we're analyzing about 15,000 global companies. The first ETF is only focused on the US, so we can go different geographies. Also, it's long only and it takes a more broader market approach. There's different strategies, geographical locations that we could utilize in the form of an ETF. As far as how the market will develop, it's been made very obvious to us that there are other players out there working on AI products or trying to utilize some form of artificial intelligence. So we do see AI products in the near future being brought more and more into the market place. I think it's important to remember that because the system is learning every day, one of the key advantages to EquBot is that our system is ahead of the learning curve. It has been learning and that we continue to use the most innovative technologies. We talked about IBM's Watson. We are also using Google DeepMind as well for some of the AI instances and we are committed to being on the most innovative platforms, whether that's Watson, Google, or some other platform in the near future.

Nate Geraci: Well Art, with that we'll have to leave it there. Really appreciate you joining us on the program today and congratulations on all the success with AIEQ. Thank you.

Art Amador: Thank you and thanks for having me.

Nate Geraci: That was Art Amador, Co-Founder of EquBot. Again, the ETF is the AI Powered Equity ETF and you can learn more about this ETF by visiting equbot.com.